Tales from the analysis pipeline: creating a catalog of vortices

The Jovian Vortex Hunter project is going strong, with over half a million classifications, so it’s a good time to talk about the science goals of the project. I want to thank all the volunteers who are on the project for helping with the classifications and Talk boards! One of the main goals of this project is to create a catalog of vortices on Jupiter, so that we can identify and compare different types of vortices based on the clouds that exist in them. Ultimately, we can then focus on individual classes of vortices, and study them in more detail!

How do we know what the different vortex types are?

This is where the second workflow (“Circle the Vortex”) comes in. Here, we ask the volunteers to circle the vortex in the subject, and identify its color. We also ask volunteers to draw ellipses around these vortices, so that we know exactly where they are within an image. So far, we have over 40,000 classifications from this workflow!

And there are so many vortices!

The panel above shows a collection of vortices and corresponding volunteer-drawn ellipses for a selection of subjects. The thin lines show the volunteer classifications and thick line is the “average” ellipse obtained by combining all of these annotations. What we find from these results is that the volunteer annotations are very precise and they are great at recognizing the different categories (different line colors show the different options — white, red, brown and dark). With these annotations, we can separate out individual vortices in the images and infer their properties (shape, color, size, etc).

What do we do with this information?

This data presents a unique opportunity since it contains very pristine information about the types of clouds in the vortex (i.e. color) and also the corresponding size, shape and location. Understanding the distribution of these vortex properties can help us better understand what is the key atmospheric process that creates these vortices. Due to the clouds, we have very little observations of what happens deeper inside Jupiter, but these storms/vortices can help us a lot by filling in the missing information!

So, let’s plot some distributions!

This is a histogram showing the distribution of vortex sizes observed by volunteers. Blue, red, brown and black correspond to white, red, brown and dark vortices respectively. What is interesting to note here is that red and brown vortices are most common at the lowest sizes (under 500km), while the most common white and dark vortices are around 1000km in sizes.

What is more interesting is the distribution of aspect ratios for the ellipse (ratio of major to minor axis). Aspect ratio close to 1 is circular while higher aspect ratios are very elliptical. Here, there’s a stark difference between the red and brown vortices compared to the white and dark vortices. The red/brown vortices have a fairly uniform distribution to high aspect ratios, showing that there are lot more of these objects that are extended.

These distributions show interesting properties of vortices and highlight real trends in how vortices are formed. For example, if we take a sample of high aspect ratio brown vortices, we get a color of barges (or barge-like vortices):

On another hand, what we can do is investigate anomalies in the distribution to study interesting vortices that are different from the rest of the distribution. For example, we find that there are a few red vortices that are larger than 5000km:

These all correspond to images of the Great Red Spot!

So, the preliminary conclusions from these plots suggest that white and dark vortices are generally circular, but much larger, while red and brown vortices can be quite small but highly elliptical! We also have anomalous vortices (large red vortices like the Great Red Spot), which warrant their own separate study.

We have vortex properties, what next?

With the calculated vortex properties, we can now create a catalog of vortices. The drawn ellipses provide a vortex center, so we can plot these back on a map to see where they form:

This is the global mosaic from perijove 18 with centers of vortices plotted back on the map. We see that there are a lot of vortices near the mid-latitudes and polar regions, but very few near the equator. This is expected from the dynamics of the jovian atmosphere — the polar regions are more turbulent and generate more vortices, compared to the equatorial regions. The Red Spot does not fit in one subject, so there are multiple centers for it across the many subjects! Let’s zoom in on the southern hemisphere:

There is good consensus on most of the vortex centers! With more classifications, we can refine the vortex centers for these vortices and generate a catalog of these features.

This is the next phase of the project

As we wind down the first workflow (we are currently in the last 4 perijoves worth of data), we will be switching over fully to the second workflow soon. As you see above, results from the second workflow are vital for interpreting more details about the physical mechanisms of vortex creation, and we still need Zooniverse’ volunteers to help us Circle these Vortices! So, if you want to help catalog these vortices and help us learn more about how they form, head on over to the Jovian Vortex Hunter project!